CEO of Patreon Jack Conte on The Future of Podcasting and AI

There are a couple different schools of thought here on what people think is going to happen in general with creators and the arts. There is the tech brain viewpoint that AI will eat up all intellectual labor, including creators and storytelling. So they are investing in AI podcast generation, AI image creation, and AI everything. And I'm not anti-AI. I actually think there are a lot of really amazing things about it. Patreon uses the tools internally. We're using it to build products faster. Our teams are doing workshops on it to learn how to leverage the technologies to build really good stuff at an even faster clip. So I don't mean for any of this to be anti-AI. But my read as someone in the tech world is that the viewpoint on it is quite singular and a bit myopic. What makes a podcaster special is not that the podcaster sounds a certain way or strings words together that make sense. Similar to what makes an artist or a musician special, a great artist will take risks with their work. They'll say something that is just on the edge of what society is willing to accept, maybe a controversial opinion, but they'll explain it in a way that is emotionally resonant with people, they'll explain why it's true, and they'll make a compelling case for that. It's interesting to watch somebody stick their neck out like that. Comedians do this all the time. They'll say something about the way the world is working right now that is really countercultural and counternarrative, but you hear them explain it and you think, "Oh, that's true." And it holds a mirror back up to us in a way that serves such an important human function in society.

Podcasters are doing that now too, though.

Podcasters are fulfilling that role. Maybe not as funny all the time, although sometimes they are funny, but they're doing that same type of thing. And I haven't yet seen a chatbot that can do that as skillfully. Maybe that is a skill they'll be able to develop. I'm skeptical. On that particular dimension, I haven't seen a model that I find as compelling as a really smart, skeptical person who is able to sense a deep societal truth and communicate it to me in a way that is socially relevant. I haven't seen that yet from AI. I've seen it in a lot of great humans that I follow. So that's one thing. Another thing is that people don't just love the work. When I listen to a Kendrick Lamar record, it's not about what the kick drum sounds like or the way the trumpet is recorded or the way the bass sounds. Those things are nice about it, but what's amazing about a Kendrick record is Kendrick. It's him. It's his pain and his purpose. I feel like he is communicating something to me. I feel like I'm learning something about him that I can draw on for myself, and I feel like a better person having listened to his record. I feel more empathetic and understanding. For my favorite artists, it's not about the output, it's about the who that's behind the output.

I think a lot of people will care about that sometimes, and I think some people won't care about it all the time. So I think there's going to be a spectrum. There will be a lot of work that is AI generated that a lot of people will consume. For example, suppose I just want information density. I want to know what happened in the world in three minutes, and I only have three minutes. Will AI be able to write an efficient paragraph summarizing what happened bullet by bullet? Probably. A series of headlines will likely be AI generated and really good. Or if I'm exercising on a treadmill and wearing a watch, AI can build a soundtrack that works with my heart rate to push me more. I'm sure some people will enjoy that soundtrack, but that doesn't mean people will stop listening to Kendrick. Music and art serve a very human, very emotional, heart-to-heart function in our lives. A lot of people will enjoy work made by other humans for a long, long time. You could say, "Jack, you're biased. You're running a platform for artists. Of course you believe that." Great. Let's check in in 10 years. I think there will be some AI generated work that people listen to and read, and it will be functional and work. And then I think there will be a lot of work that continues to be deeply human and more about understanding another person and their struggles, helping me as a listener or viewer or reader feel less alone in my own humanity. That type of work will exist for a long, long time.

That's art.

I'm not. So here's the important thing. I'm not out there saying kill all AI. AI is bad. I think I started by saying the platforms bifurcate humanity into two camps on every issue. They polarize us. They filter us into camps. This has happened with AI, and I think that's really sad because a lot of us are using the tools but are now almost too afraid to even talk about how we're using them. I'm seeing conversations in private that are different from those in public. In public, everyone is kind of saying, "Oh, fuck AI," and yet I'm using ChatGPT three times a week to help me do research. This is what I mean by the algorithm sorting us into two camps. It has made it really untenable to have reasonable conversations about AI. The truth is, I think there are some really interesting uses for AI. It can be really helpful for people. And I also think there are a lot of really dangerous things about it that we need to be really careful of and cognizant of. We saw what social media did to humanity over its first chapter. Let's not fucking do that again with AI. Why would we do that again? Let's be more careful. Let's be more deliberate. Let's not let it strip away the income from artists like social media did the first time around. And by the way, that's happening again with AI. The models have absorbed the collective work of humanity and not compensated creators for any of that work. When you go to a model and say, "Make me an image that looks like Studio Ghibli," it outputs a Studio Ghibli image, and Studio Ghibli collects nothing for that. They don't get paid a cent. That's fucked up. And that needs to change. And it's possible for that to be true at the same time as it's true that these models have some real utility.

And that's also true. So let's talk about it and work through it and figure out what's good and dangerous and not good. I wish people could discuss nuance. And you very much have to join one of the two sides on social media, otherwise you're eviscerated. I think that's such a dangerous place to be as a society.